There is broad agreement that the U.S. health care system fails to meet the needs of the most rapidly growing segment of the U.S. population - the oldest old. Public reporting and pay for performance initiatives are already underway to address the uneven quality of the current delivery system. There is growing concern, however, that current efforts could harm the frail elderly by focusing on measuring adherence to clinical guidelines developed for younger populations. These guidelines promote interventions that lead to long-term mortality benefit but may engender short-term risk. The shortened survival of the very old or frail may mean they accrue the risks without benefit. We believe that valid measures of performance in health care for the oldest old can be created using patient-centered values, rather than disease-based indicators, as the conceptual framework of measurement. As a junior investigator, my long term goal is to develop a research career devoted to developing, testing, and implementing innovative approaches to measuring health care system quality for the oldest old. The goals of this proposal are to build the necessary skills and the pilot data to become an independent investigator in this field. The specific career development aims are: (1) to develop expertise in assessment of values or care needs of the oldest old;(2) to develop advanced expertise in the use, integration, and analysis of survey data and a diverse group of administrative claims including Medicare claims and the Minimum Data Set;(3) to refine skills in statistical analysis and modeling needed for the development and application of performance measures. The research goal of this proposal is to develop and pilot test, in two health systems that differ in aggressiveness of end-of-life care, quality measures that reflect patient-centered values in health care for the oldest old using survey and claims-based methods. The specific research aims are: (1) to develop a survey instrument that measures support for individualized choices for people older than 80 years of age or their caregivers;(2) to develop claims-based indicators of patient-valued processes of care in terms of patient experience (ex. continuity) and clinical services (nursing home time and end-of-life care);and (3) to pilot test the multi-modal quality measures in two specific health system and explore hypotheses about the relationships between support for patient-centered decisions and continuity with clinical services among the very old. Relevance: Billions of dollars are spent on providing acute and long-term care to the very old. The projected quadrupling of this population creates a public health imperative to provide the highest quality care possible at the lowest cost. The goal of this proposal is to create the measures of quality that are needed to monitor and improve the quality of care for the very old.